On 04 Okt 2003, Paul Watson <paul@oninit.com> wrote:
> OS: Solaris 9
>
> Everything is working fine, no problems as far I can tell then
> all of a sudden our monitoring picks up a read cache drop and
> I get the following onstat -p. The DBserver is for a website,
> traffic levels haven't changed, and there are no background
> jobs, and I'm seeing very little disk activity but the cache
> still reads 0%. The stats have not be cleared down
>
> Any ideas
>
> informix: onstat -p
>
> Informix Dynamic Server Version 9.40.UC1 -- On-Line -- Up 17
> days 22:08:25 -- 2493440 Kbytes
>
> Profile
> dskreads pagreads bufreads %cached
> 194114421 312676058 177660467 0.00
apparently the no. of bufreads rolled over the 32 bit limit (~4G).
The definition of %cached (see the admin's guide) says that it is
set to zero, if dskreads is greater than bufreads.